[WMS] Re: Thoughts on standards

  • From: "Sunir Shah" <sunir@xxxxxxxxx>
  • To: <wiki@xxxxxxxxxxxxx>
  • Date: Wed, 6 Oct 2004 13:06:51 -0400

I don't think of WikiSyntax is not like a programming 
language at all. I can see why programmers look at it 
like that, but that isn't the best practice, I think.

It's best to ignore the concept of wiki syntax and
look at it from the point of view of a user. From their
point of view, they just want to bash text into the wiki
and have it work. The wiki's job is to make the text they
bash into the wiki look good, and 'do what they mean'. 

The purpose of the syntax description is a script-oriented
view. When that viewpoint finds itself into the user's
mindset, it is jarring and annoying. Rather, wikis are good
because if you can type and drool on the 'Save' button, they
do something.

The reason why the syntaxes are divergent are that 
divergent user communities have different conventions
and needs. If you paste in e-mails to a wiki, the colon
syntax for indentation that UseMod employs is not the
best; rather, a greater sign would be better since that
is the quotation syntax employed by most e-mail clients.
If you write a lot of academic papers in your wiki, LaTeX 
is important.

The goal of the script is to facilitate the bashing in
of text. The more it impedes that process by creating 
rules that users have to learn before they use the script,
the less writerly the medium becomes and the more difficult
it gets.

To that end, the syntax employed in the script has to fit
a particular context. The syntax we want to standardize will
not be the syntax used to play Go games on SenseisWiki. 
Rather, we should aim to standardize syntax in particular

If you look at how the web has evolved, once HTML hit 4.0
with CSS, DHTML, and JavaScript (ECMAScript), it became too
expensive to publish to the web. When blogs became easy
to publish to, people flocked to them, even if their 
information architecture is not so great. Before, however,
the web exploded because it was so easy to bash text into
a .html file and publish it so others can read it. Now, it
is less easy, if even by social convention where web designers
fall over themselves to make simple things more complex with
CSS. XHTML will make things even more difficult and arcane.
Indeed, I would say XHTML is an example of engineers wresting
control over the printing press from the publisher, and that's
not a valuable end goal.

All technical standards once founded create an environment 
for feature creep as they are 'versioned' up to include more
and more scope. XHTML, CSS, DHTML, JS are examples of this
happening to the very simple HTML1.0 which was used to be
a simple syntax to do the kinds of stuff they were writing
at CERN (more or less).

I suggest we forcibly limit scope on this effort to layers
or modules, and the layer we focus on initially is the syntax
used to do simple print-based formatting.

Ordered Lists
Undered Lists
Paragraph breaks

And let's face it, a) the best way to do that level of formatting
is often a WYSIWYG editor in many user domains, b) that doesn't
really seem to matter given that there are an abdundanace of markup 
standards that already define those basic featuresets, eg. 
reStructuredText, XHTML-Basic, etc. 

So perhaps what we're talking about is redundant, but then maybe
we can just keep things simple since we don't need to define the 
one syntax to rule the world.


Other related posts: